Soft chitosan microbeads scaffold for 3D functional ......Soft chitosan microbeads scaffold for 3D...

lable at ScienceDirect

Biomaterials 156 (2018) 159e171

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Soft chitosan microbeads scaffold for 3D functional neuronal networks

Maria Teresa Tedesco a, 1, Donatella Di Lisa a, 1, Paolo Massobrio a, Nicol�o Colistra a,Mattia Pesce b, Tiziano Catelani c, Elena Dellacasa a, Roberto Raiteri a, d,Sergio Martinoia a, d, Laura Pastorino a, *

a University of Genova, Dept. of Informatics, Bioengineering, Robotics and System Engineering, Via Opera Pia 13, 16145, Genova, Italyb Nanoscopy and Nikon Centre, Istituto Italiano di Tecnologia, via Morego, 30, 16163, Genova, Italyc Electron Microscopy Facility, Istituto Italiano di Tecnologia, via Morego, 30, 16163, Genova, Italyd CNR - Institute of Biophysics, Via De Marini, 6, 16149, Genova, Italy

a r t i c l e i n f o

Article history:Received 31 July 2017Received in revised form15 November 2017Accepted 27 November 2017Available online 28 November 2017

Keywords:ChitosanMicrobeadsNeuronal culture3D networkMicro-electrode arrays (MEAs)

* Corresponding author.E-mail address: [email protected] (L. Pastor

1 M.T. and D.D.L. contributed equally to this work.

https://doi.org/10.1016/j.biomaterials.2017.11.0430142-9612/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

The availability of 3D biomimetic in vitro neuronal networks of mammalian neurons represents a pivotalstep for the development of brain-on-a-chip experimental models to study neuronal (dys)functions andparticularly neuronal connectivity. The use of hydrogel-based scaffolds for 3D cell cultures has beenextensively studied in the last years. However, limited work on biomimetic 3D neuronal cultures hasbeen carried out to date. In this respect, here we investigated the use of a widely popular polysaccharide,chitosan (CHI), for the fabrication of a microbead based 3D scaffold to be coupled to primary neuronalcells. CHI microbeads were characterized by optical and atomic force microscopies. The cell/scaffoldinteraction was deeply characterized by transmission electron microscopy and by immunocytochemistryusing confocal microscopy. Finally, a preliminary electrophysiological characterization by micro-electrode arrays was carried out.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

The physico-chemical characteristics of the extracellular matrix(ECM) play a fundamental role in regulating relevant physiologicalcellular processes and in different pathological situations [1,2].Consequently, it has become clear that cellular organization in 3D iscrucial to study biological functions [3,4]. To date, most in vitrofunctional studies have been performed using oversimplifiedtraditional monolayer cultures. However, in the last few years, agrowing number of research groups have been focusing on thesetting up of cellular models which mimic the in vivo microenvi-ronment at a higher extent [5,6]. This approach This approach hasproven to be essential to gain information on pathological pro-cesses like cancer, where cell-cell and cell-microenvironment in-teractions play a major role [7]. The availability of 3D cultureplatforms, specifically designed to mimic different tissues towardsthe development of organ-on-a-chip [8], is expected to have astrong impact not only in the study of physiological and

ino).

pathological processes, but also in drug screening and in toxicityassays [3,9e13]. In the process of developing 3D in vitro models, afundamental step is represented by the engineering and tailoring ofa 3D matrix containing adequate chemical and mechanical signalsin order to support the cell phenotypes of interest. In this respect,synthetic and natural hydrogels have been used to develop 3Dmodels for soft tissues, because of hydrophilicity, biocompatibility,biodegradability, and tunable microporosity [11,14e21]. Amongsoft tissue models, 3D interconnected networks of neuronal cellsare very useful to investigate in a reduced in vitro model, neuronal(dys)functions and connectivity for applications ranging from basicneuroscience to drug screening [22e24].

Under this perspective, we have recently demonstrated that 3Dhippocampal networks, made by self-assembled glass microbeadsas scaffold, and coupled to micro-electrode arrays (MEAs), repre-sent a suitable in vitro model for neurophysiological studies alter-native and complementary to the classical 2D neuronal networkmodels [25]. Matrix stiffness and composition are the most criticalproperties, since they can influence growth dynamics, synapticdensity, and electrophysiological activity of the neuronal network[26e28]. In this work, we propose the use of soft porous hydrogelmicrobeads platform, mimicking the physico-chemical character-istics of the ECM, for the growth of 3D neuronal networks. Natural

mailto:[email protected]://crossmark.crossref.org/dialog/?doi=10.1016/j.biomaterials.2017.11.043&domain=pdfwww.sciencedirect.com/science/journal/01429612http://www.elsevier.com/locate/biomaterialshttps://doi.org/10.1016/j.biomaterials.2017.11.043https://doi.org/10.1016/j.biomaterials.2017.11.043https://doi.org/10.1016/j.biomaterials.2017.11.043

M.T. Tedesco et al. / Biomaterials 156 (2018) 159e171160

soft materials used in 3D neuronal cell culture have been mainlybased on the use of proteins, such as collagen [29,30], or of ECMprotein mixtures, such as Matrigel™ matrix [6,31].

In the last few years, polysaccharides have been proposed andwidely used as biomimetic materials for scaffold fabrication for ahuge variety of cell types [32,33]. However, the use of poly-saccharides for 3D neuron cultures has been quite limited ascompared to other cell types. Among the polysaccharides, algi-nate, hyaluronic acid, and gellan gum have been mainly investi-gated [34e37]. Very few studies have addressed the use ofchitosan, and the electrophysiological behavior of chitosan based3D neuronal cultures is still unknown [38e46]. Chitosan is acopolymer of glucosamine and N-acetyl-glucosamine, obtained bythe deacetylation of chitin, which is the main component ofcrustacean and insect exoskeletons [47]. Chitosan can behave as apolycation under acidic conditions (pH < 6), due to the proton-ation of free amino groups [48]. This polysaccharide is well knownfor its biocompatibility, biodegradability, muco-adhesiveness, andantibacterial and antifungal activity [49]. Interestingly, as previ-ously demonstrated, chitosan enhances neuron attachment,proliferation and neurite extension, and exerts a potent neuro-protective action [50,51]. The aim of this study was to explore theuse of chitosan as scaffold material for 3D neuronal networkscoupled with MEAs.

As a first step, 2D physically cross-linked chitosan films wereprepared by phase inversion (liquid to solid) in an ethanol/so-dium hydroxide solution [52] and their interaction with neuronswas assessed by inverted microscopy, with and without treat-ment with adhesion proteins. The 2D cultures carried out ontochitosan films were investigated only in the view of gaininginformation on the bioactivity of chitosan in terms of celladhesion ad network development. Physically cross-linked chi-tosan microbeads were then fabricated by an aerodynamically-assisted jetting technique, characterized by optical and atomicforce microscopies (AFM) and then used as scaffold for 3D hip-pocampal neuron cultures. The 3D neuronal networks werecharacterized morphologically by transmission electron micro-scopy (TEM), by immunofluorescence techniques and 3Dimaging with a confocal microscope. The spontaneous electro-physiological activities of the obtained 3D networks wererecorded after 21 days of in vitro culture (21 DIV); results werecompared with those obtained using glass microbeads [25] as 3Dscaffold for neuronal growth.

2. Materials and methods

2.1. Preparation of chitosan films

Chitosan (CHI, mediummolecular weight, 75e85% deacetylated,code 448877, lot MKBD4275V, from Pandalus Borealis), ethanol,sodium hydroxide and acetic acid were purchased from Sigma-Aldrich.

CHI was dissolved in 0.1 M acetic acid under continuous stirringfor 2 h and filtered through a syringe filter (5 mm) to remove anyundissolved material. Films were prepared from CHI solutions atconcentrations 1% and 2% w/v. CHI solutions (1 ml) were poured ona petri dish (∅35 mm) and allowed to dry before exposing them to1 ml of gelling solution overnight. The gelling solution was pre-pared by mixing H2Odd 40%, Ethanol 60% and NaOH 2% w/v. Thegelling solution compositionwas optimized on the basis of the datapresent in the literature and on results obtained by the FT-IRcharacterization of CHI samples crosslinked at different concen-trations of NaOH (data not shown) [53,54]. The obtained films werewashed several times with distilled water.

2.2. Preparation of chitosan microbeads

Two CHI concentrations were tested for microbeads fabrication,namely 1% and 2% w/v. 3 ml of filtered CHI solutions were extrudedusing a microencapsulation unit (Nisco Encapsulation Unit VARJ30) equipped with a conical nozzle having a diameter of 0.25 mm[52]. The extrusion flow rate was 0.4 ml/min under 100 mbarpressure for CHI 1%, whereas for CHI 2% the extrusion flow rate was0.5 ml/min under 200 mbar pressure. The generated micro-droplets were collected into 150 ml of gelling solution bath whilecontinuously stirring at 200 rpm. The distance from the nozzle tothe gelling solution was set at 6 cm. The resulting microbeads wereleft in contact with the gelling solution for 30 min at room tem-perature to ensure complete solidification. Afterwards, the gellingsolutionwas removed through centrifugation (1000 rpm for 5min),followed by four washing steps in distilled water. The productionyield was evaluated using optical microscopy. An inverted opticalmicroscope (IX-51 Olympus microscope equipped with a DP70digital camera and with a 10� N.A. 0.25 PhC objective) was used totake images of the microbeads. From the collected images, afterbinarization, the “analyze particles tool” of ImageJ software (NIH,USA) was used to evaluate the projected areas of the microbead.Areas of particles touching each other were separated bywatershedsegmentation.

2.3. Characterization of chitosan microbeads

The mean particle size and size distributions were evaluatedusing optical microscopy. The mean particle size was evaluatedusing ImageJ software as described above.

2.3.1. Water content2.5 � 106 microbeads in 0.5 ml of water were weighed inside an

Eppendorf tube and were then lyophilized. The weight of the driedmicroparticles was measured and the water content was calculatedusing the following equation, where Ws is the weight of hydratedmicrobeads and Wd is the weight of the lyophilized ones:

Water content ¼ ðWs � WdÞWs

� 100%

2.3.2. Atomic force microscopy (AFM)A commercial atomic force microscope, equipped with a closed

loop scanner capable of 9 mmvertical range (Keysight Technologies,model 5500ILM), was used tomeasure both the topography and thestiffness of the microbeads. Rectangular micro-cantilevers (Mikro-mash HQ:CSC38, type B, nominal spring constant k ¼ 0.03 N/m)either with a conical tip or without any tip were employed. Imagesof the topography of single beads were obtained in contact mode,by careful adjusting the lowest possible force to keep the contactduring the whole scan. In order to evaluate the stiffness of thebeads, standard force curves were recorded and the region aftercontact was considered for further analysis. The applied load forcantilever deflections was calculated by first converting the outputvoltage, from the AFM four-segment photodetector, into nanome-ters of deflection, and then by multiplying the deflection by thecantilever spring constant. The conversion factor was calculated bytaking several force curves onto a hard glass substrate each time thelaser spot on the cantilever had to be adjusted, and by consideringthe reciprocal of the average slope of the constant compliance re-gion of the curves. When using sharp conical tips, the load versusindentation curve was evaluated to extract the elastic modulus ofthe sample using the model proposed by Oliver and Pharr [55] as

M.T. Tedesco et al. / Biomaterials 156 (2018) 159e171 161

already described in Ref. [56]. When using tipless cantilevers theload versus microbead deformation curvewas evaluated in order toextract the stiffness of the bead in analogy with unconfinedcompression testing. All measurements were performed at a con-stant approaching/retracting speed of 1 mm/s. This allowed us tocompare results, despite the viscous (i.e. speed-dependent)response of the CHI microbeads. In order to take into accountintra-sample heterogeneity, 16 � 16 ¼ 256 force curves wererecorded over a regular grid over a 5 � 5 mm.

Microbeads were adsorbed onto the surface of a petri dish pre-modified by the deposition of a layer of polyethylenimine (PEI1 mg/ml in pure water, from Sigma Aldrich), followed by a layer ofpolystyrene sulfonate (PSS 2 mg/ml, Sigma-Aldrich). For bothsamples, three maps of 16 � 16 curves were collected onto threedifferent microbeads randomly selected over the petri surface.

2.4. Cell preparation

Hippocampi were dissected and removed from embryonicSprague-Dawley rats at gestational day 18 under sterile condi-tions. Hippocampal fetal tissue was enzymatically digested inTrypsin 0.125% in Caþþ and Mgþþ free Hank's (Gibco Invitrogen)for 200 at 37 �C. The enzymatic process was quenched by addingculture medium supplemented with 10% of FBS (Sigma-Aldrich)then the tissuewas mechanically dissociated with a smoothly fire-polished Pasteur pipette. Neurons were re-suspended in platingmedium consisting of Neurobasal medium (Gibco Invitrogen)with 2% w/v B-27 Supplement (Gibco Invitrogen), 1% Glutamax(Gibco Invitrogen), 1% Pen-Strepto (Gibco Invitrogen). Cultureswere maintained in incubator at 37 �C in a 5% CO2, 95% humidityatmosphere for 3e4 weeks by replacing half of the medium once aweek [57]. The experimental protocol was approved by the Eu-ropean Animal Care Legislation (2010/63/EU), by the ItalianMinistry of Health in accordance with the D.L. 116/1992 and by theguidelines of the University of Genova. All efforts were made toreduce the number of animals used for the project and to mini-mize their suffering [58].

2.4.1. Preparation of 2D networks on CHI films and 3D networks onCHI microbeads

The day before plating, CHI films andmicrobeads were sterilizedby exposure to ethanol 70% for 2 h. The sterilized samples werethen washed with sterile water 5 times, normalized in cell culturemedium and used for the cell culture experiments. To evaluate thebioaffinity between CHI and neurons, cultures were prepared usingboth films and microbeads treated and untreated with adhesionproteins (a.p.).

Fig. 1. Set-up configuration: (A) Micro-electrode arrays (MEAs) made up of 60 planar microsquare grid with inserted PDMS (internal diameter 5 mm) constraint on the active area; (B

In the first case, both films and microbeads were exposed to amix of a. p., namely Laminin:P-D-Lysine (1:1), at the concentrationof 0.05 mg/ml in sterile water (L-2020; P-6407 Sigma-Aldrich) andleft in the incubator overnight at 37 �C. The a. p. were then washedaway from the films with sterile water, while the microbeads werecentrifuged three times, for 5 min at 1000 rpm. Each centrifugationstep was followed by a washing step in sterile water. In the secondcase, both films and microbeads were sterilized and used withoutany further treatment. Before cell plating, both films andmicrobeads were washed in a Neurobasal medium. Hippocampalneurons were plated onto the film surface at a seeding concentra-tion of 1000 cell/ml with a final cell density of 800 cell/mm2. In thecase of cell plating onto the microbead surface, the microbeadswere exposed to the cell suspension in complete Neurobasal me-dium; the ratio of the number of microbeads to the number ofneuronswas nominally 1:4. Eppendorf vials were used for this step;106 microbeads/ml and 4 � 106 cells/ml were mixed and after aninterval of around 3e4 h they aggregated and formed small clus-ters. The vials were kept in horizontal position and turned for12e16 times at 20e25 min intervals in order to expose the wholemicrobead surface to the suspended cells. At the end of theincubation-adhesion phase, the neuron-microbead aggregateswere left to deposit slowly at the bottom of the vial. Finally, theywere carefully collected with a micropipette in small volumes(30e35 ml), and directly transferred onto standard petri dishes(∅35 mm) for subsequent immunocytochemistry characterization,or plated onto the MEA surface for electrophysiologicalcharacterization.

The day before plating, MEAs were assembled with donuts-shaped Poly-dimethyl-siloxane (PDMS) structures (internal andexternal diameters: 5 and 22 mm respectively, height: 650 mm) toconfine the self-assembled microbeads and neurons onto a circularsurface of ~20 mm2 around the active electrodes area (Fig. 1A).MEAs (assembled as explained above) were sterilized in the oven at120� for 2 h. At the end of the sterilization process, the chips weretreated only on the area delimited by the PDMS structure, with amix of a. p. namely Laminin: P-D-Lysine (1:1), at the concentrationof 0.05 mg/ml in sterile water (L-2020; P-6407 Sigma) and left inthe incubator overnight at 37 �C. The coating solutionwas removedfrom the MEAwhich was then washed twice with water and left todry under the laminar hood until the plating took place. Similarly towhat performed in Ref. [25], hippocampal neurons withoutmicrobeads were first plated onto the MEA surface to create a firstmonolayer of cells at a final concentration of 800e1000 cell/mm2.3e4 h after plating, 30e35 ml of neuron-microbead aggregates weretransferred inside the PDMS confinement structure onto the areaon which hippocampal neurons were previously seeded (Fig. 1B).

electrodes (TiN/SiN, 30 mm electrode diameter, 200 mm spaced) arranged over an 8 � 8) 3D CHI scaffold macroscale assembly onto MEA labeled for MAP-2.


Around 4.5 � 104 microbeads and 1.5 � 105 cells were transferredinto MEAs.

2.5. Morphological characterization of 3D neuronal networks bytransmission electron microscopy

In order to analyze the samples with Transmission ElectronMicroscopy (TEM), the 3D networks on CHI 2% were fixed for 2 h ina fixative solution (2% Glutaraldehyde, in buffer Na-Cacodylate0.1 M) and then post-fixed (2 h) in a solution 1% OsO4, 1,5% Po-tassium Hexacyanoferrate, in Na-cacodylate buffer 0.1 M. Subse-quently, theywere stained overnight in a 1% Uranyl acetate aqueoussolution and dehydrated with series of alcohols. TEM samples wereinfiltrated with Propylene Oxide and low viscosity Spurr resin (SPI-Chem). Once the resin hardened, 70 nm thick sections were cutwith a Leica EMU C6 ultra-microtome. TEM images were collectedby means of Jeol JEM 1011 (Jeol, Japan) TEM, operating at an ac-celeration voltage of 100 kV, and recorded with a 11 Mp fiber op-tical charge-coupled device (CCD) camera (Gatan Orius SC-1000).All used reagents were from Sigma-Aldrich.

2.6. Morphological characterization of neuronal networks byimmunocytochemistry

To assess the expression of specific neuronal markers, hippo-campal cultures were fixed in 4% paraformaldehyde in phosphatebuffer solution (PBS), pH 7.4 for 30 min at room temperature.Permeabilization was achieved with PBS containing 0.5% Triton-X100 for 15 min at room temperature and non-specific binding ofantibodies was blocked with an incubation of 45 min in a blockingbuffer solution consisted of PBS, 0.3% BSA (bovine serum albuminSigma) and 0.5% FBS. Cultures were incubated with primary anti-body diluted in PBS Blocking buffer for 2 h at room temperature orincubated at 4 �C overnight in a humidified atmosphere. Cultureswere rinsed three times with PBS and finally exposed to the sec-ondary antibodies. The following primary antibodies were used forCHI films: MAP-2 1:500 (monoclonal or polyclonal Synaptic Sys-tem), TUBULIN bIII, clone TU-20 (similar TUJ1) 1:500 (ChemiconMillipore), NeuN 1:200 (Chemicon millipore), VGAT and VGLUT11:500 (Synaptic System), Synapsin 1:200 (Synaptic System), Dapi1:10000 (Sigma). The following primary antibodies were used forCHI microbeads: MAP-2 1:500 (monoclonal or polyclonal SynapticSystem), TUBULIN bIII, clone TU-20 (similar TUJ1) 1:500 (ChemiconMillipore), NeuN 1:200 (Chemicon Millipore), Dapi 1:10000(Sigma). To verify the presence of glial cells in the culture, we fixedand exposed to the marker GFAP 1:500 (CHI microbeads) or 1:1000(CHI films) monoclonal or polyclonal antibodies (Sigma). Cultureswere rinsed twice with PBS and finally exposed to the secondaryantibodies: Alexa Fluor 488, Alexa Fluor 549, Alexa Fluor 633 Goatanti mouse or Goat anti rabbit, diluted 1:700 and 1:1000 (Invi-trogen Life Technologies S. Donato Milanese).

To observe the perineuronal net-like structure, we exposedsamples to Wisteria floribunda 1:200 (Sigma-Aldrich) as primaryantibody for 24 h and Streptavidin Alexa Fluor 488, 1:700 (Invi-trogen Life Technologies S. Donato Milanese) for 6 h as secondaryantibody.

A table of all used antibodies and respective dilutions is reportedin Supplementary materials (Table S1).

2.6.1. Optical microscopy and confocal imagingAn inverted IX-51 Olympus microscope equipped with a DP70

digital camera coupled with CPlan 10� N.A. 0.25 PhC objective wasused to acquire contrast phase images of CHI microbeads coupledwith neurons. An Olympus BX-51 upright microscope was used forimmunofluorescence evaluation of the biological samples and the

image acquisition was done with a Hamamatsu Orca ER II digitalcooled CCD camera driven by Image ProPlus software (MediaCybernetic).

Confocal imaging was acquired on two different microscopes:Leica TCS SP5 AOBS TandemDMI6000 invertedmicroscope coupledwith objective Leica IRAPO 25�, 0.95 NA (Leica Microsystems,Mannheim, Germany) and Leica TCS SP5 AOBS Tandem DM6000upright microscope coupled with objective Leica IRAPO 25�, 0.95NA (Leica Microsystems Srl, Italy). Data were analyzed by means ofthe LASX V2.0 software (Leica Microsystems Srl, Italy).

2.7. MEA recording and analysis

The spontaneous electrophysiological activity of 3D hippocam-pal neuronal networks was recorded at 21e24 days in vitro (DIV) bymeans of micro-electrode arrays (MEAs) made up of 60 planarmicroelectrodes (TiN/SiN, 30 mm electrode diameter, 200 mmspaced) arranged over an 8 � 8 square grid (except the four elec-trodes at the corners), supplied by Multi Channel Systems (MCS,Reutlingen, Germany). The electrophysiological activity was ac-quired with the 2100 System (MEA 2100-System, MCS), and signalswere sampled at 10 kHz. Recordings were performed for 30 minoutside the incubator at a temperature of 37 �C. To prevent evap-oration and changes of the pH medium, a slow flow of humidifiedgas (5% CO2, 20% O2, 75% N2) was constantly delivered during themeasurement sessions into a small plastic box covering theexperimental MEA setup.

2.7.1. Data and statistical analysisData analysis was performed by using a custom software pack-

age named SPYCODE [59], developed in MATLAB (The Mathworks,Natick, MA, USA). Spike detection was performed by using thePrecise Timing Spike Detection (PTSD) algorithm [60]. The algo-rithm requires three parameters: a different threshold set to 8times the standard deviation of the baseline noise, a peak lifetimeperiod (set at 2 ms) and a refractory period (set at 1 ms). To char-acterize the electrophysiological activity, we extracted some firstorder statistics. In particular, we evaluated the mean firing rate(MFR), i.e., the number of spikes per second of each channel and thepercentage of random spikes, i.e., the fraction of spikes outsidebursts. We also performed burst detection according to the methoddescribed in Ref. [61]. A burst is a sequence of spikes having an ISI(inter-spike interval, i.e., time intervals between consecutivespikes) smaller than a reference value (set at 100 ms in our ex-periments), and containing at least a minimum number ofconsecutive spikes (set at 5 spikes). The parameters extracted fromthis analysis are the mean bursting rate (MBR) and the mean burstduration (MBD), which are the frequency and the duration of thebursts at the single channel level respectively. The same approachused for the detection of bursts was applied for the detection ofquasi synchronous events at network level called network bursts[59]. The extracted parameters are the network bursting rate (NBR)and the network burst duration (NBD). NBR computes the number ofnetwork bursts per minute, while NBD is the temporal extension ofthese events.

Statistical analysis was carried out using OriginPro 8 (OriginLabCorporation, Northampton, MA, USA). All data are presented asmean ± standard error of the mean. Statistical analysis was per-formed using a non-parametric Kruskal-Wallis test, since data donot follow a normal distribution (evaluated by the Kolmogorov-Smirnov normality test). Differences were considered statisticallysignificant when p < 10�3. In order to determine which of thesample pairs are significantly different, post-hoc test, using Dunn'stest, has been applied.


3. Results

3.1. Preparation and characterization of CHI microbeads

CHI films were investigated only in view of gaining informationon the bioactivity of CHI itself. Physically cross-linked CHI filmswere thus prepared onto petri dishes and the growth of a 2Dneuronal network on top of CHI films was characterized. Relevantresults are discussed in section 3.2.

1% and 2% CHI microbeads were prepared and characterized inview of their use as scaffolds for neuronal growth. The instrumentalparameters, for microbeads production, were optimized in order topromote the formation of the micro-droplet spray and to avoidaggregation of microbeads on the air/gelling solution interface. Theproduction yields were evaluated to be around 0.9 � 106 and0.7 � 106 per batch for 1% and 2% CHI respectively.

Optical microscopy images of the obtained samples were ac-quired and analyzed. The results indicated a spherical shape and asize ranging from 40 to 90 mm, with an average diameter of66 ± 20 mm, for CHI 1%, while for CHI 2% the size ranged from 40 to160 mm, with an average diameter of 100 ± 40 mm, Fig. 2AeB.Watercontent values were found to be 98.4% and 99.3% for 2% and 1% CHImicrobeads, respectively.

AFM topography showed nanometer sized features onto arounded profile. The elastic modulus of the microbeads was firstevaluated by AFM indentation measurements using microcanti-levers with conical tips. The elastic modulus measured on 2%CHI microbeads was in the range 15e25 kPa, whereas 1% chi-tosan microbeads were too soft to reliably determine the pointof contact and, thus, to calculate the elastic modulus. Therefore,

Fig. 2. Characterization of CHI microbeads: (A) Histogram of the distribution of CHI 1%Topography AFM image (second order flattened) of 15 � 15 mm2 of a single 2% CHI microbeadline); (D) Stiffness values measured by AFM on 1% and 2% CHI microbeads. (For interpretatversion of this article.)

we used a tipless cantilever of the same type to press against asingle microbead. The slope of the force curve after contactresulted constant for a wide range of applied forces (1e10 nN)and with negligible hysteresis between loading and unloading(Supplementary Materials Fig. S1). This slope represents thestiffness of the microbeads. Stiffness values obtained from theconstant compliance region of curves performed on differentmicrobeads using the same cantilever and the same approach-retract speed can be directly compared. In Fig. 2D averagestiffness values measured on 1% and 2% CHI microbeads areplotted. The values are normalized versus the average stiffnessof 2% CHI beads. 1% microbeads were found to be, on average, 18fold softer than the average stiffness of the probed 2%microbeads. Interestingly enough, the range of the elasticityvalue for 1% CHI microbeads that can be inferred by our mea-surements (1/18 of 15e25 kPa) falls in the same range of re-ported elasticity values for brain tissue (0.7e1 kPa) [62,63] (seealso Supplementary Materials).

3.2. Preparation and characterization of 2D neuronal networks onCHI films

As a first step, the bioaffinity of CHI towards neurons wascharacterized. To conduct this study, a simplified standard 2D cul-ture model was adopted and cells were thus plated onto the surfaceof CHI films. In order to evaluate the ability of physically cross-linked CHI to promote neuronal adhesion and development, filmsboth untreated and treated with a. p. were used. Fig. 3 shows theimages of neuronal networks developing onto (A) 2% CHI untreatedfilm, (B) 2% CHI film treated with a. p. and (C) petri dish untreated

microbeads size; (B) Histogram of the distribution of CHI 2% microbeads size; (C)in culture medium, the insert shows the profile of the raw data from a single line (blueion of the references to colour in this figure legend, the reader is referred to the web

Fig. 3. Optical contrast phase images of 2D neuronal network: (A) 2% CHI film untreated with a. p. at DIV 15; (B) 2% CHI film treated with a. p at DIV 15; (C) petri dish untreatedwith a. p. at DIV 15; (D) 2% CHI untreated film labeled for Tubulin-bIII (green) and NeuN (red) at DIV 25; Scale bar: 50 mm. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)


with a. p at DIV 15 and (D) 2% CHI untreated film labeled forTubulin-bIII and NeuN at DIV 25.

With respect to the morphological development of the neuronalnetwork, it can be observed that cell morphology is similar whencells are plated onto untreated and treated 2% CHI films (Fig. 3A andB). Same results were obtained for cells cultured onto 1% CHI films(data not shown). In Fig. 3C, considered as the negative control,neurons, as expected, tended to form clusters and no network wasobtained. Cells showed a homogeneous distribution and the for-mation of a dense network onto 2% CHI untreated film even at DIV25 (Fig. 3D). Moreover, the presence of functional structures at DIV25 was evaluated by 2D networks on 2% CHI film labeled for Syn-apsin and VGAT-VGLUT (Supplementary Materials Fig. S2). Theseresults confirm that CHI naturally promotes adhesion, neuritegrowth and structural development of the network, even withoutany treatment with a. p. Overall, both treated and untreated filmswere able to sustain the growth and development of cells overthree-four weeks, with the formation of a stable network.

In order to quantify the composition of the cellular populationduring the in vitro network development on untreated film, a per-centage variation of both the neuronal and glial population hasbeen evaluated (Supplementary Materials Fig. S3).

3.3. Preparation and characterization of 3D neuronal networks onCHI microbeads

As a first step, in order to verify the bioaffinity of CHI also in theform of microbeads, cells were cultured onto CHI microbeads bothuntreated and treated with a. p. The obtained 3D cultures wereobserved by contrast phase optical microscopy during the first twoweeks of culture in vital conditions. In both cases, a branched and

entangled expression of neurites and the presence of healthyneurons, which was demonstrated by the refractivity of theneuronal soma, were observed.

All subsequent experiments were carried out onto 3D neuronalnetworks grown onto 1% and 2% CHI microbeads pre-treated witha. p. This was done in order to compare the properties of the 3Dnetworks grown onto CHI microbeads with the ones grown ontoglass microbeads as described in Ref. [25]. TEM characterizationwas carried out in order to appreciate the interaction between CHImicrobeads and cultured cells. From low magnification imaging(Fig. 4A) it is clear that cells and microbeads (marked with blackasterisks) create a dense network. Because of the electron mi-croscopy staining, cell bodies and dendrites appear darkercompared to the microbeads, allowing to observe that cells bothenvelop and penetrate the chitosan scaffold. Higher magnificationimages show more in detail the interaction among neurons andCHI microbeads. Fig. 4B shows a neuronal cell and its axonpushing out between two beads, meeting then the dendrites fromanother neuron. Besides this, it is clear how smaller dendritesenter inside the chitosan, as underlined by the black arrows. InFig. 4C is reported a detail of the interface between cells and CHImicrobeads, where many neurons components are evident: cellbodies, axons with distinct tubulin cytoskeleton, small dendrites,spines and synapses. From these two images, it is clear that on themicrobeads surface there is a dense network made up of axonsand dendrites of different size, while only the smaller dendritespenetrate into the CHI microbeads. In Fig. 4D is reported a detail ofa dendrite taken far from the microbead surface (about 15 mm).The size of these terminations is smaller (

Fig. 4. Low-mag TEMmicrograph of a portion of chitosan scaffold with the neuronal network: (A) CHI-microbeads are marked with the asterisks, while cells appear darker, thedark arrow highlight a glial cell. (B) A neuron grown between two microbeads (1000�). The dark arrows indicate dendrites inside the chitosan. (C) A detail (2500�) of the interfacecell-CHI microbeads. Synapses are marked with the letter s. (D) High-mag (15000�) detail of dendrites taken far (15 mm) from the bead surface.


Indirect immunofluorescence techniques were then used toassess the in vitromorphological and functional cell behavior and tocharacterize the 3D structure of the network. To this purpose, after25 DIVs, at the end of the recording sessions (see section 3.3), 3Dcell-scaffolds were fixed with PFA 4% and labeled by using apanel of ad-hoc selected antibody molecules. The 3D neural net-works were then characterized by confocal microscopy on MAP2labeled neurons (Supplementary Movie 1) and on MAP-2 andTubulin bIII (Supplementary Movie 2). Overall, the thickness of the3D neuronal network on the 2% CHI microbeads was evaluated tobe around 300e500 mm.

Supplementary video related to this article can be found athttps://doi.org/10.1016/j.biomaterials.2017.11.043.

Fig. 5 shows the neuronal network development around the CHImicrobeads (A, C) and around glass microbeads (B).

We can see (Fig. 5A, left and middle) neuronal soma fromwhichrich neuritic arborizations depart. This is particularly evidentaround the CHI microbead surfaces while it becomes partiallyfragmented due to the penetration of neurites into themicroporousvolume of the microbeads (white arrows) (Supplementary Movie 3and Movie 4). In the case of the glass microbeads, the neuronalnetwork development was confined onto the surface of microbe-ads, without any fragmentation (Fig. 5B, left). Moreover, the shapeof the soma was found to be spherical in both cases (Fig. 5 AeB,middle), as the one observed in vivo [25,64]. Fig. 5A (right) shows asection of the 3D culture where it is possible to observe the closeassembly between CHI microbeads mediated by neuronal cells.Instead, in the case of glass microbeads a hexagonal structure wasobserved characterized by well-defined and separated microbeads(Fig. 5B, right). Fig. 5C left shows the structural proteins of thecytoskeleton of the 3D network on 2% CHI microbeads. Fig. 5C rightshows the high density of synaptic puncta present on the 3Dnetwork.


In the formation of brain-like constructs a pivotal role in thesurvival and differentiation of neurons is played by glial cells [65].In order to highlight the morphology of glia cells, fixed 3D and 2Dcultures were exposed to GFAP primary antibody followed by sec-ondary antibody Alexa Fluor 549.

It can be observed that the GFAP positive cells cultured both on2D film and 3D microbeads 2% CHI (Fig. 6AeB) present a differentmorphology compared to GFAP positive cells cultured at the sameconditions but on the 2D petri dish surface (Fig. 6C).

These results suggest that the chemical and mechanical envi-ronments play a relevant role on the morphological behavior.Similar observations were already reported in previous works[23,66e70].

The 3D structure of neuronal networks after 24 days of culturefixed and immunolabeled for the dendritic marker MAP-2 is illus-trated in Fig. 7. Both the 3D reconstruction of 148 mm z-stack of thehippocampal network and the projections along different axes areshown, thus giving a comprehensive view of the neuronal networkonto the CHI microbeads scaffold (Fig. 7B). The max intensity pro-jection of the orthogonal view is represented in this figure: XZprojection (Fig. 7A) shows CHI microbeads profiles wrapped byneuronal network; Fig. 7C and D shows XY e YZ projections.

3.4. Functional characterization of 3D networks

In order to perform the electrophysiological characterization ofthe 3D networks grown onto CHI microbeads and to compare theobtained results with the ones reported in Ref. [25], CHI microbeadswere pre-treated with a. p., then mixed with neurons and finallyplated ontoMEA. As reported in Ref. [25], before this final step, a 2Dneuronal network was directly coupled to the active area of MEA inorder to establish a good communication between the 3D cultureand the underlying microelectrodes.

Fig. 8A shows the spontaneous activity (raw signal) of 10 s of a 3DCHI network as recorded from onemicroelectrode and characterized

https://doi.org/10.1016/j.biomaterials.2017.11.043https://doi.org/10.1016/j.biomaterials.2017.11.043

Fig. 5. Confocal microscope images of 3D neural network at DIV 25: (A) 3D neural network on 2% CHI microbeads (left), single 2% CHI microbead surrounded by almost sixneurons (middle) and a section of 3D neural network on 2% CHI microbeads (right), MAP-2 (green) and Synapsin (red). (B) 3D neuronal network on glass microbeads (left), singleglass microbead surrounded by five neurons (middle) and a section of 3D neural network on glass microbeads (right), MAP-2 (green and red). The blue arrows point the cell somawhile the white one points neuritic fragmentation. (C) 3D neural network on 2% CHI microbeads (left) labeled for MAP-2 (green), Tubulin bIII (red) and DAPI (blu), 3D neuralnetwork on 2% CHI microbeads (right) labeled for Synapsin (green) and DAPI (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to theweb version of this article.)

Fig. 6. Optical images of glial cells labeled for GFAP (DIV 25): (A) 2D network on CHI film; (B) 3D network on 2% CHI microbeads; (C) 2D network on petri dish. The white arrowspoint CHI microbeads.


Fig. 7. Max intensity projection Orthogonal View and Rendering 3D of population neurons network on the scaffold labeled (DIV 24) for MAP-2: Cross-sectioning alongdifferent axes XZ (A), XY(C), YZ (D); (B) Volumetric representation (XYZ) of neuronal networks from the same sample.


by two bursts and random spikes. The global electrophysiologicalbehavior of representative 3D networks, is qualitatively showed inthe raster plots of Fig. 8B and C, where 300 s of spontaneous activityare displayed. In both experimental conditions (CHI, Fig. 8B and glassmicrobeads, Fig. 8C), quasi-synchronous network bursts (NB) aremixedwith random spiking activity. However, 3D networks with CHImicrobeads scaffold exhibit a global activity characterized by longerbursts than glass microbeads ones. After 21 DIV, we recorded 30minof spontaneous activity of n ¼ 3 CHI 1% networks, and n ¼ 3 CHI 2%networks, and we compared the obtained results to n ¼ 3 cultureswhere the 3D scaffold was realized by means of glass microbeads(SupplementaryMaterials Table S2). Fig. 8DeI shows the parametersextracted from the analyzed spike data. CHI 1% networks presentedvalues ofMFR (2.3± 0.14 spikes/s), statistically different from the CHI2% ones (0.86± 0.05 spikes/s; p< 0.001) but similar to theMFRof the3D glass microbeads networks (2.97 ± 0.55 spikes/s). All the 3Dexperimental configurations display high values of random spikingactivity (Fig. 8E): specifically, CHI 1% and 2% networks show higher(statistically significant) values with respect to glass microbeadsones. Regarding the bursting behavior, the MBR of CHI 2% networksshowed the lowest value (3.74 ± 0.31 (bursts/min)) which is signif-icantly different (p < 0.001) from the CHI 1% and glass microbeadsnetworks. On the other way, round (Fig. 8G), CHI 1% networksexhibited a MBD (310.5 ± 18.19 (ms)) significantly higher (p < 0.001)than the other two configurations that share similar MBD values(175.10 ± 16.41 ms and 190.70 ± 10.53 ms for CHI 2% and glassmicrobeads).

Finally, the NB activity was investigated by computing thenetwork mean bursting rate (NBR; Fig. 8H) and duration (NBD;Fig. 8I). NBR was similar for CHI 1% and glass microbeads networksand statistically different with respect to CHI 2% networks. Again,

for the NBD, CHI 1%, was different from CHI 2% and glassmicrobeads that are in turn characterized by shorter bursts(0.32 ± 0.036 s and 0.43 ± 0.067 s, respectively).

4. Discussion

4.1. Characterization of CHI microbeads

In this work, we explored the use of CHI microbeads to activelysupport 3D functional neuronal cultures. CHI was chosen for itsbiocompatibility, biodegradability, and low cost [6]. Moreover, inthe literature it is reported that the positive charges of primaryamines onto the polymer backbone favor the electrostatic inter-action with the negatively charged cell membranes [42,44,71,72],promoting cell adhesion and growth.

In general, the stiffness, porosity, and electrostatic charge ofthe scaffold concur in neuritic development and extension. In ourcase, the stiffness of 1% CHI microbeads was found to be compa-rable to that of brain tissue (Fig. 2D). A difference in the stiffness,among 1% and 2% CHI microbeads, was observed and could beattributed to the increase in the concentration of CHI, corre-sponding to an increase in the density of the polymeric chains.Therefore, higher ionic interactions between the CHI chains seemto be quite predictable, as the concentration of chitosan increasedfrom 1% to 2% [73]. Besides, as shown by water content result,microbeads with lower concentration of chitosan have a highercontent of water which consequently caused a decrease in stiff-ness [74] However, this effect is not a major factor because thedifference of water content in the two samples is not remarkableand it is logic to state that the increase in the ionic interactions isthe main mechanism for the stiffness growth.

Fig. 8. Spontaneous activity characterization: (A) 10 s of raw data recorded from a single microelectrode. Raster plot showing 300 s of spontaneous activity of 3D network on (B)CHI 1% microbeads and(C) on glass bead; (D) mean firing rate, (E) percentage of random spiking activity, (F) mean bursting rate, (G) mean burst duration, (H) network bursting rate,(I) network bursts duration. (Kruskal-Wallis, *p � 10�3).


Finally, AFM revealed darker areas on the surface, which mightrepresent holes whose apparent dimensions are in agreement withdata obtained by TEM (see Section 4.3). (Fig. 2C). These character-istics combined with bioaffinity of CHI, due to the presence ofprimary amines, contributed to the formation of a dense neuronalnetwork onto CHI microbeads.

4.2. Characterization of 2D neuronal networks on CHI films

In the effort of investigating the intrinsic bioaffinity of CHI, wefirstly characterized its ability to induce cell attachment and neuriteoutgrowth without any pre-treatment with a. p. As a first step, thischaracterization was carried out using standard 2D cell culture


models onto the surface of CHI films. Quite surprisingly, non-pre-treated CHI films were able to support neuronal growth during aperiod of more than 15 days.

A similar observation was previously done for soft alginatehydrogels, which were able to support neural cell cultures inmonolayer or spheroids [75,76]. However, the use of alginate assupporting material for neuronal cultures is controversial, since itrequires Ca2þ ions for its ionic cross-linking and it is well knownthat neurons and glial cells are extremely sensitive to Ca2þ ions,even at nM concentrations [77].

To our knowledge, this is the first work reporting the ability ofpure CHI to support neuronal cell attachment and functionalneuronal network development. This represents a valuablecontribution in the search for low-cost biomimetic culture systems,which can have important applications in neuropharmacology,toxicology, and regenerative medicine [78,79].

4.3. Characterization of 3D neuronal networks on CHI microbeads

The experimental design of this study was partially inspired bythe results of a previous work in which a scaffold made by glassmicrobeads was used [25]. Indeed, to allow a direct comparison ofresults deriving from the use of CHI and glass microbeads, CHImicrobeads were pre-treated with a. p. and employed as 3D sup-port for cell attachment and growth.

The immunocytochemistry and confocal microscopy character-ization allowed us to gain information on (i) the morphology of the3D structure of mature neuronal networks after 24 days of cultureand (ii) the distinct features of the two cell populations dissectedfrom hippocampal rat brain tissues, namely neurons and glial cells.

Regarding cell morphology, the neuronal somata were found tobe round, like the ones observed in the brain tissue. The ability ofCHI microbeads to maintain the in vivo cell morphology wasalready reported by Garcia-Giralt et al., who studied their interac-tion with human chondrocytes [80].

This result underlines that the combination of different factors,including substrate stiffness, 3D arrangement, and chemical cuesaltogether contribute to support an in vivo-like growth of theneuronal network.

The scaffold topography, characterized by confocal microscopy,showed that while the micro-scale dendritic extensions weredistributed on the external surface of the CHI microbeads, thenano-scale ones tended to penetrate the hydrogel, contributing tothe formation of a compact structure.

This speculation was confirmed by TEM analysis, which allowedto understand in depth the micro- and nano-structure of theneuron-microbeads assembly. TEM micrographs (Fig. 4) clearlysupport the data obtained by confocal microscopy, putting in evi-dence that CHI microbeads are enveloped in a dense network ofneural dendrites and axons. At the same time, we had the evidencethat smaller dendrites are allowed to enter and spread inside theCHI microbeads, proving its porosity to neural dendrites.

The astrocyte glial fraction also proliferates on CHI microbeadsand its morphology was again similar to the one found in braintissue (i.e., having a thin morphology and expression in GFAP). Thisobservation was already reported by others [67,68] for 3D in vitrocultures, thus suggesting a substrate-induced morphologicaldependence. Interestingly, also the morphology of glial cellscultured onto CHI film was found to be stretched (Fig. 6).

In themeanwhile, glial cells and the natural ECM, spontaneouslyproduced by the neurons network in culture, were responsible forthe assembly of the microbeads after four-weeks in culture. In or-der to verify the presence of natural ECM, we assessed the forma-tion of perineuronal net-like structures in our culture systems usingWFA (Supplementary Movie 5).


Finally, it is worth mentioning that no evident differences be-tween the microbeads made by CHI at 1% compared to those at 2%were observed in terms of neuronal morphology and distribution ofthe biological material on the scaffold.

4.4. Functional characterization of 3D neuronal networks on CHImicrobeads

Regarding the electrophysiological characterization, after 21DIVs, the 3D neuronal networks developed onto CHI microbeadspresented electrophysiological patterns similar to the onesobserved for the glass microbeads in terms of the percentage ofrandom spiking and bursting behavior. As already observed byFrega et al. [25] in the case of glass microbeads, the percentage ofrandom spiking of the 3D CHI networks presents higher values thanthose observed in 2D cultures (Supplementary materials Fig. S4).Here we found a further increase of random spiking with respect tothe 3D glass microbead networks. Moreover, we observed that 3DCHI 1% networks show more synchronous bursts (MBR, NBR), withrespect to 3D CHI 2% with associated longer duration for bothbursts and network bursts. This activity indicates the formation of avery dense network with a high degree of connectivity as alsosuggested by the immunostaining for Synapsin (Fig. 5C right) andfor MAP-2 (Fig. 7). Therefore, the 3D CHI model presents possibleadvantages that would merit further investigations: (i) stiffnesssimilar to the living brain tissue; (ii) no need for pre-treatmentwitha. p.; (iii) high-level of connectivity; (iv) in vivo like electrophysi-ological behavior.

From the other side, it should be considered that the numberof active electrodes was significantly lower in the case of CHImicrobeads based 3D cultures. This was due to a lack of contactsbetween MAE surface and the overhanging 3D assembly. Neu-rons, cultured on the MEA surface, were partially transferredfrom the 2D monolayer to the surface of the overhanging 3Dassembly. This was probably due to the higher bioaffinity of CHIthan MEA surface. Moreover, it should be considered that weobserved a stable assembly between CHI microbeads and cellsonly after the first week in culture. Therefore, the mechanicalstresses caused by replacements of the medium in the first weekof culture might have contributed to the weakening of the 3Dassembly.

Arrays of 3D microelectrodes ad hoc designed would provideeasier physical integration with the culture and more resolved ac-cess to the electrophysiological network activity.

5. Conclusions

Chitosan microbeads based scaffolds were specifically opti-mized and adapted in order to be integrated onto planar MEAs tostudy and better understand the functional properties of bio-mimetic 3D hippocampal networks. Chitosan microbeads bothtreated and untreated with adhesion factors were tested andboth of them proved to be reliable supports, able to sustain theneuronal population during the growth in a 3D space. At thesame time, the chitosan microbeads guaranteed both amorphological and structural development of a functionalnetwork. Finally, we demonstrated that the neuronal networkitself was responsible for the assembly and the stabilization ofthe 3D chitosan based structure. In conclusion, CHI seems to be apromising scaffolding-support for developing 3D neuronal net-works towards the design and implementation of brain-on-a-chip microsystems.

https://doi.org/10.1016/j.biomaterials.2017.11.043


Funding

This research did not receive any specific grant from fundingagencies in the public, commercial, or not-for-profit sectors.

Acknowledgements

Maria Concetta Miniaci for her valuable comments on this workand Giorgio Carlini for the technical support.

Appendix A. Supplementary data

Supplementary data related to this article can be found athttps://doi.org/10.1016/j.biomaterials.2017.11.043.

References

[1] T. Dvir, et al., Nanotechnological strategies for engineering complex tissues,Nat. Nanotechnol. 6 (1) (2011) 13e22.

[2] E. Cukierman, et al., Taking cell-matrix adhesions to the third dimension,Science 294 (5547) (2001) 1708e1712.

[3] B.A. Justice, N.A. Badr, R.A. Felder, 3D cell culture opens new dimensions incell-based assays, Drug Discov. Today 14 (1e2) (2009) 102e107.

[4] L.G. Griffith, M.A. Swartz, Capturing complex 3D tissue physiology in vitro,Nat. Rev. Mol. Cell Biol. 7 (3) (2006) 211e224.

[5] F. Pampaloni, E.G. Reynaud, E.H. Stelzer, The third dimension bridges the gapbetween cell culture and live tissue, Nat. Rev. Mol. Cell Biol. 8 (10) (2007)839e845.

[6] D.B. Kolesky, et al., 3D bioprinting of vascularized, heterogeneous cell-ladentissue constructs, Adv. Mater. 26 (19) (2014) 3124e3130.

[7] K.H. Benam, et al., Engineered in vitro disease models, Annu. Rev. Pathol. 10(2015) 195e262.

[8] D. Huh, et al., Microfabrication of human organs-on-chips, Nat. Protoc. 8 (11)(2013) 2135e2157.

[9] B. Weigelt, C.M. Ghajar, M.J. Bissell, The need for complex 3D culture modelsto unravel novel pathways and identify accurate biomarkers in breast cancer,Adv. Drug Deliv. Rev. 69e70 (2014) 42e51.

[10] J. Lee, et al., In vitro toxicity testing of nanoparticles in 3D cell culture, Small 5(10) (2009) 1213e1221.

[11] M.W. Tibbitt, K.S. Anseth, Hydrogels as extracellular matrix mimics for 3D cellculture, Biotechnol. Bioeng. 103 (4) (2009) 655e663.

[12] S. Breslin, L. O'Driscoll, Three-dimensional cell culture: the missing link indrug discovery, Drug Discov. Today 18 (5e6) (2013) 240e249.

[13] A. Roth, T. Singer, The application of 3D cell models to support drug safetyassessment: opportunities & challenges, Adv. Drug Deliv. Rev. 69e70 (2014)179e189.

[14] M.J. Mahoney, K.S. Anseth, Three-dimensional growth and function of neuraltissue in degradable polyethylene glycol hydrogels, Biomaterials 27 (10)(2006) 2265e2274.

[15] L.S. Wang, et al., Injectable biodegradable hydrogels with tunable me-chanical properties for the stimulation of neurogenesic differentiation ofhuman mesenchymal stem cells in 3D culture, Biomaterials 31 (6) (2010)1148e1157.

[16] A. Kunze, et al., Micropatterning neural cell cultures in 3D with a multi-layered scaffold, Biomaterials 32 (8) (2011) 2088e2098.

[17] D.R. Nisbet, et al., Neural tissue engineering of the CNS using hydrogels,J. Biomed. Mater. Res. Part B Appl. Biomater. 87 (1) (2008) 251e263.

[18] S.R. Caliari, J.A. Burdick, A practical guide to hydrogels for cell culture, Nat.Methods 13 (5) (2016) 405e414.

[19] A.S. Hoffman, Hydrogels for biomedical applications, Ann. N. Y. Acad. Sci. 944(2001) 62e73.

[20] Y. Du, et al., Directed assembly of cell-laden microgels for fabrication of 3Dtissue constructs, Proc. Natl. Acad. Sci. U. S. A. 105 (28) (2008) 9522e9527.

[21] C.M. Magin, D.L. Alge, K.S. Anseth, Bio-inspired 3D microenvironments: a newdimension in tissue engineering, Biomed. Mater. 11 (2) (2016) 022001.

[22] S. Pautot, C. Wyart, E.Y. Isacoff, Colloid-guided assembly of oriented 3Dneuronal networks, Nat. Methods 5 (8) (2008) 735e740.

[23] T.B. Puschmann, et al., A novel method for three-dimensional culture ofcentral nervous system neurons, Tissue Eng. Part C Methods 20 (6) (2014)485e492.

[24] S. Bosi, et al., From 2D to 3D: novel nanostructured scaffolds to investigatesignalling in reconstructed neuronal networks, Sci. Rep. 5 (2015) 9562.

[25] M. Frega, et al., Network dynamics of 3D engineered neuronal cultures: a newexperimental model for in-vitro electrophysiology, Sci. Rep. 4 (2014) 5489.

[26] J. Lantoine, et al., Matrix stiffness modulates formation and activity ofneuronal networks of controlled architectures, Biomaterials 89 (2016) 14e24.

[27] I. Levental, P.C. Georges, P.A. Janmey, Soft biological materials and theirimpact on cell function, Soft Matter. 3 (3) (2007) 299e306.

[28] A.P. Balgude, et al., Agarose gel stiffness determines rate of DRG neuriteextension in 3D cultures, Biomaterials 22 (10) (2001) 1077e1084.

[29] M. Antman-Passig, O. Shefi, Remote magnetic orientation of 3D collagenhydrogels for directed neuronal regeneration, Nano Lett. 16 (4) (2016)2567e2573.

[30] D. Ge, et al., Culture and differentiation of rat neural stem/progenitor cells in athree-dimensional collagen scaffold, Appl. Biochem. Biotechnol. 170 (2)(2013) 406e419.

[31] M. Uemura, et al., Matrigel supports survival and neuronal differentiation ofgrafted embryonic stem cell-derived neural precursor cells, J. Neurosci. Res.88 (3) (2010) 542e551.

[32] J.P. Frampton, et al., Fabrication and optimization of alginate hydrogel con-structs for use in 3D neural cell culture, Biomed. Mater 6 (1) (2011) 015002.

[33] N. Broguiere, L. Isenmann, M. Zenobi-Wong, Novel enzymatically cross-linkedhyaluronan hydrogels support the formation of 3D neuronal networks, Bio-materials 99 (2016) 47e55.

[34] M. Matyash, et al., Swelling and mechanical properties of alginate hydrogelswith respect to promotion of neural growth, Tissue Eng. Part C Methods 20 (5)(2014) 401e411.

[35] K.E. Crompton, et al., Polylysine-functionalised thermoresponsive chitosanhydrogel for neural tissue engineering, Biomaterials 28 (3) (2007) 441e449.

[36] PattersonJ. et al., Biomimetic materials in tissue engineering. Mater. Today, 13(1),14e22;

[37] J.A. Hunt, et al., Hydrogels for tissue engineering and regenerative medicine, J. Mater. Chem.B 2 (33) (2014) 5319e5338.

[38] R. Lozano, et al., 3D printing of layered brain-like structures using peptidemodified gellan gum substrates, Biomaterials 67 (2015) 264e273.

[39] N.D. Leipzig, M.S. Shoichet, The effect of substrate stiffness on adult neuralstem cell behavior, Biomaterials 30 (36) (2009) 6867e6878.

[40] S. Wrobel, et al., In vitro evaluation of cell-seeded chitosan films for peripheralnerve tissue engineering, Tissue Eng. Part A 20 (17e18) (2014) 2339e2349.

[41] N.B. Skop, et al., Optimizing a multifunctional microsphere scaffold to improveneural precursor cell transplantation for traumatic brain injury repair, J. TissueEng. Regen. Med. 10 (10) (2016) E419eE432.

[42] C.M. Valmikinathan, et al., Photocrosslinkable chitosan based hydrogels forneural tissue engineering, Soft Matter. 8 (6) (2012) 1964e1976.

[43] S. Guan, et al., Chitosan/gelatin porous scaffolds containing hyaluronic acidand heparan sulfate for neural tissue engineering, J. Biomater. Sci. Polym. Ed.24 (8) (2013) 999e1014.

[44] Z. Cao, R.J. Gilbert, W. He, Simple agarose-chitosan gel composite system forenhanced neuronal growth in three dimensions, Biomacromolecules 10 (10)(2009) 2954e2959.

[45] C.Y. Sung, et al., Probing neural cell behaviors through micro-/nano-patternedchitosan substrates, Biofabrication 7 (4) (2015) 045007.

[46] C. Huang, et al., The migration and differentiation of hUC-MSCs(CXCR4/GFP)encapsulated in BDNF/chitosan scaffolds for brain tissue engineering, Bio-med. Mater. 11 (3) (2016) 035004.

[47] M. Rinaudo, Chitin and chitosan: properties and applications, Prog. Polym. Sci.31 (7) (2006) 603e632.

[48] T. Kean, M. Thanou, Biodegradation, biodistribution and toxicity of chitosan,Adv. Drug. Deliv. Rev. 62 (1) (2010) 3e11.

[49] A. Anitha, et al., Chitin and chitosan in selected biomedical applications, Prog.Polym. Sci. 39 (9) (2014) 1644e1667.

[50] J.M. Zuidema, et al., Fabrication and characterization of tunable poly-saccharide hydrogel blends for neural repair, Acta Biomater. 7 (4) (2011)1634e1643.

[51] M. Dash, et al., Chitosanda versatile semi-synthetic polymer in biomedicalapplications, Prog. Polym. Sci. 36 (8) (2011) 981e1014.

[52] C.A. Cust�odio, et al., Functionalized microparticles producing scaffolds incombination with cells, Adv. Funct. Mater. 24 (10) (2014) 1391e1400.

[53] Q. He, et al., Preparation of chitosan films using different neutralizing solu-tions to improve endothelial cell compatibility, J. Mater. Sci. Mater. Med. 22(12) (2011) 2791e2802.

[54] J. Du, et al., Comparative evaluation of chitosan, cellulose acetate, and poly-ethersulfone nanofiber scaffolds for neural differentiation, Carbohydr. Polym.99 (2014) 483e490.

[55] G.M. Pharr, W.C. Oliver, F.R. Brotzen, On the generality of the relationshipamong contact stiffness, contact area, and elastic modulus during indentation,J. Mater. Res. 7 (3) (2011) 613e617.

[56] E. Defranchi, et al., Imaging and elasticity measurements of the sarcolemma offully differentiated skeletal muscle fibres, Microsc. Res. Tech. 67 (1) (2005)27e35.

[57] G.J. Brewer, C.W. Cotman, NMDA receptor regulation of neuronal morphologyin cultured hippocampal neurons, Neurosci. Lett. 99 (1989) 268e273.

[58] M. Tedesco, et al., Interfacing 3D engineered neuronal cultures to micro-electrode arrays: an innovative in vitro experimental model, JoVE (104)(2015) e53080.

[59] L.L. Bologna, et al., Investigating neuronal activity by SPYCODE multi-channeldata analyzer, Neural Netw. 23 (6) (2010) 685e697.

[60] A. Maccione, et al., A novel algorithm for precise identification of spikes inextracellularly recorded neuronal signals, J. Neurosci. Methods 177 (1) (2009)241e249.

[61] V. Pasquale, S. Martinoia, M. Chiappalone, A self-adapting approach for thedetection of bursts and network bursts in neuronal cultures, J. Comput.Neurosci. 29 (1) (2010) 213e229.

[62] A.J. Engler, et al., Matrix elasticity directs stem cell lineage specification, Cell126 (4) (2006) 677e689.

https://doi.org/10.1016/j.biomaterials.2017.11.043http://refhub.elsevier.com/S0142-9612(17)30773-1/sref1http://refhub.elsevier.com/S0142-9612(17)30773-1/sref1http://refhub.elsevier.com/S0142-9612(17)30773-1/sref1http://refhub.elsevier.com/S0142-9612(17)30773-1/sref2http://refhub.elsevier.com/S0142-9612(17)30773-1/sref2http://refhub.elsevier.com/S0142-9612(17)30773-1/sref2http://refhub.elsevier.com/S0142-9612(17)30773-1/sref3http://refhub.elsevier.com/S0142-9612(17)30773-1/sref3http://refhub.elsevier.com/S0142-9612(17)30773-1/sref3http://refhub.elsevier.com/S0142-9612(17)30773-1/sref3http://refhub.elsevier.com/S0142-9612(17)30773-1/sref4http://refhub.elsevier.com/S0142-9612(17)30773-1/sref4http://refhub.elsevier.com/S0142-9612(17)30773-1/sref4http://refhub.elsevier.com/S0142-9612(17)30773-1/sref5http://refhub.elsevier.com/S0142-9612(17)30773-1/sref5http://refhub.elsevier.com/S0142-9612(17)30773-1/sref5http://refhub.elsevier.com/S0142-9612(17)30773-1/sref5http://refhub.elsevier.com/S0142-9612(17)30773-1/sref6http://refhub.elsevier.com/S0142-9612(17)30773-1/sref6http://refhub.elsevier.com/S0142-9612(17)30773-1/sref6http://refhub.elsevier.com/S0142-9612(17)30773-1/sref7http://refhub.elsevier.com/S0142-9612(17)30773-1/sref7http://refhub.elsevier.com/S0142-9612(17)30773-1/sref7http://refhub.elsevier.com/S0142-9612(17)30773-1/sref8http://refhub.elsevier.com/S0142-9612(17)30773-1/sref8http://refhub.elsevier.com/S0142-9612(17)30773-1/sref8http://refhub.elsevier.com/S0142-9612(17)30773-1/sref9http://refhub.elsevier.com/S0142-9612(17)30773-1/sref9http://refhub.elsevier.com/S0142-9612(17)30773-1/sref9http://refhub.elsevier.com/S0142-9612(17)30773-1/sref9http://refhub.elsevier.com/S0142-9612(17)30773-1/sref9http://refhub.elsevier.com/S0142-9612(17)30773-1/sref10http://refhub.elsevier.com/S0142-9612(17)30773-1/sref10http://refhub.elsevier.com/S0142-9612(17)30773-1/sref10http://refhub.elsevier.com/S0142-9612(17)30773-1/sref11http://refhub.elsevier.com/S0142-9612(17)30773-1/sref11http://refhub.elsevier.com/S0142-9612(17)30773-1/sref11http://refhub.elsevier.com/S0142-9612(17)30773-1/sref12http://refhub.elsevier.com/S0142-9612(17)30773-1/sref12http://refhub.elsevier.com/S0142-9612(17)30773-1/sref12http://refhub.elsevier.com/S0142-9612(17)30773-1/sref12http://refhub.elsevier.com/S0142-9612(17)30773-1/sref13http://refhub.elsevier.com/S0142-9612(17)30773-1/sref13http://refhub.elsevier.com/S0142-9612(17)30773-1/sref13http://refhub.elsevier.com/S0142-9612(17)30773-1/sref13http://refhub.elsevier.com/S0142-9612(17)30773-1/sref13http://refhub.elsevier.com/S0142-9612(17)30773-1/sref13http://refhub.elsevier.com/S0142-9612(17)30773-1/sref14http://refhub.elsevier.com/S0142-9612(17)30773-1/sref14http://refhub.elsevier.com/S0142-9612(17)30773-1/sref14http://refhub.elsevier.com/S0142-9612(17)30773-1/sref14http://refhub.elsevier.com/S0142-9612(17)30773-1/sref15http://refhub.elsevier.com/S0142-9612(17)30773-1/sref15http://refhub.elsevier.com/S0142-9612(17)30773-1/sref15http://refhub.elsevier.com/S0142-9612(17)30773-1/sref15http://refhub.elsevier.com/S0142-9612(17)30773-1/sref15http://refhub.elsevier.com/S0142-9612(17)30773-1/sref16http://refhub.elsevier.com/S0142-9612(17)30773-1/sref16http://refhub.elsevier.com/S0142-9612(17)30773-1/sref16http://refhub.elsevier.com/S0142-9612(17)30773-1/sref17http://refhub.elsevier.com/S0142-9612(17)30773-1/sref17http://refhub.elsevier.com/S0142-9612(17)30773-1/sref17http://refhub.elsevier.com/S0142-9612(17)30773-1/sref18http://refhub.elsevier.com/S0142-9612(17)30773-1/sref18http://refhub.elsevier.com/S0142-9612(17)30773-1/sref18http://refhub.elsevier.com/S0142-9612(17)30773-1/sref19http://refhub.elsevier.com/S0142-9612(17)30773-1/sref19http://refhub.elsevier.com/S0142-9612(17)30773-1/sref19http://refhub.elsevier.com/S0142-9612(17)30773-1/sref20http://refhub.elsevier.com/S0142-9612(17)30773-1/sref20http://refhub.elsevier.com/S0142-9612(17)30773-1/sref20http://refhub.elsevier.com/S0142-9612(17)30773-1/sref21http://refhub.elsevier.com/S0142-9612(17)30773-1/sref21http://refhub.elsevier.com/S0142-9612(17)30773-1/sref22http://refhub.elsevier.com/S0142-9612(17)30773-1/sref22http://refhub.elsevier.com/S0142-9612(17)30773-1/sref22http://refhub.elsevier.com/S0142-9612(17)30773-1/sref23http://refhub.elsevier.com/S0142-9612(17)30773-1/sref23http://refhub.elsevier.com/S0142-9612(17)30773-1/sref23http://refhub.elsevier.com/S0142-9612(17)30773-1/sref23http://refhub.elsevier.com/S0142-9612(17)30773-1/sref24http://refhub.elsevier.com/S0142-9612(17)30773-1/sref24http://refhub.elsevier.com/S0142-9612(17)30773-1/sref25http://refhub.elsevier.com/S0142-9612(17)30773-1/sref25http://refhub.elsevier.com/S0142-9612(17)30773-1/sref26http://refhub.elsevier.com/S0142-9612(17)30773-1/sref26http://refhub.elsevier.com/S0142-9612(17)30773-1/sref26http://refhub.elsevier.com/S0142-9612(17)30773-1/sref27http://refhub.elsevier.com/S0142-9612(17)30773-1/sref27http://refhub.elsevier.com/S0142-9612(17)30773-1/sref27http://refhub.elsevier.com/S0142-9612(17)30773-1/sref28http://refhub.elsevier.com/S0142-9612(17)30773-1/sref28http://refhub.elsevier.com/S0142-9612(17)30773-1/sref28http://refhub.elsevier.com/S0142-9612(17)30773-1/sref29http://refhub.elsevier.com/S0142-9612(17)30773-1/sref29http://refhub.elsevier.com/S0142-9612(17)30773-1/sref29http://refhub.elsevier.com/S0142-9612(17)30773-1/sref29http://refhub.elsevier.com/S0142-9612(17)30773-1/sref30http://refhub.elsevier.com/S0142-9612(17)30773-1/sref30http://refhub.elsevier.com/S0142-9612(17)30773-1/sref30http://refhub.elsevier.com/S0142-9612(17)30773-1/sref30http://refhub.elsevier.com/S0142-9612(17)30773-1/sref31http://refhub.elsevier.com/S0142-9612(17)30773-1/sref31http://refhub.elsevier.com/S0142-9612(17)30773-1/sref31http://refhub.elsevier.com/S0142-9612(17)30773-1/sref31http://refhub.elsevier.com/S0142-9612(17)30773-1/sref32http://refhub.elsevier.com/S0142-9612(17)30773-1/sref32http://refhub.elsevier.com/S0142-9612(17)30773-1/sref33http://refhub.elsevier.com/S0142-9612(17)30773-1/sref33http://refhub.elsevier.com/S0142-9612(17)30773-1/sref33http://refhub.elsevier.com/S0142-9612(17)30773-1/sref33http://refhub.elsevier.com/S0142-9612(17)30773-1/sref34http://refhub.elsevier.com/S0142-9612(17)30773-1/sref34http://refhub.elsevier.com/S0142-9612(17)30773-1/sref34http://refhub.elsevier.com/S0142-9612(17)30773-1/sref34http://refhub.elsevier.com/S0142-9612(17)30773-1/sref35http://refhub.elsevier.com/S0142-9612(17)30773-1/sref35http://refhub.elsevier.com/S0142-9612(17)30773-1/sref35http://refhub.elsevier.com/S0142-9612(17)30773-1/sref37http://refhub.elsevier.com/S0142-9612(17)30773-1/sref37http://refhub.elsevier.com/S0142-9612(17)30773-1/sref37http://refhub.elsevier.com/S0142-9612(17)30773-1/sref37http://refhub.elsevier.com/S0142-9612(17)30773-1/sref38http://refhub.elsevier.com/S0142-9612(17)30773-1/sref38http://refhub.elsevier.com/S0142-9612(17)30773-1/sref38http://refhub.elsevier.com/S0142-9612(17)30773-1/sref39http://refhub.elsevier.com/S0142-9612(17)30773-1/sref39http://refhub.elsevier.com/S0142-9612(17)30773-1/sref39http://refhub.elsevier.com/S0142-9612(17)30773-1/sref40http://refhub.elsevier.com/S0142-9612(17)30773-1/sref40http://refhub.elsevier.com/S0142-9612(17)30773-1/sref40http://refhub.elsevier.com/S0142-9612(17)30773-1/sref40http://refhub.elsevier.com/S0142-9612(17)30773-1/sref41http://refhub.elsevier.com/S0142-9612(17)30773-1/sref41http://refhub.elsevier.com/S0142-9612(17)30773-1/sref41http://refhub.elsevier.com/S0142-9612(17)30773-1/sref41http://refhub.elsevier.com/S0142-9612(17)30773-1/sref42http://refhub.elsevier.com/S0142-9612(17)30773-1/sref42http://refhub.elsevier.com/S0142-9612(17)30773-1/sref42http://refhub.elsevier.com/S0142-9612(17)30773-1/sref43http://refhub.elsevier.com/S0142-9612(17)30773-1/sref43http://refhub.elsevier.com/S0142-9612(17)30773-1/sref43http://refhub.elsevier.com/S0142-9612(17)30773-1/sref43http://refhub.elsevier.com/S0142-9612(17)30773-1/sref44http://refhub.elsevier.com/S0142-9612(17)30773-1/sref44http://refhub.elsevier.com/S0142-9612(17)30773-1/sref44http://refhub.elsevier.com/S0142-9612(17)30773-1/sref44http://refhub.elsevier.com/S0142-9612(17)30773-1/sref45http://refhub.elsevier.com/S0142-9612(17)30773-1/sref45http://refhub.elsevier.com/S0142-9612(17)30773-1/sref46http://refhub.elsevier.com/S0142-9612(17)30773-1/sref46http://refhub.elsevier.com/S0142-9612(17)30773-1/sref46http://refhub.elsevier.com/S0142-9612(17)30773-1/sref47http://refhub.elsevier.com/S0142-9612(17)30773-1/sref47http://refhub.elsevier.com/S0142-9612(17)30773-1/sref47http://refhub.elsevier.com/S0142-9612(17)30773-1/sref48http://refhub.elsevier.com/S0142-9612(17)30773-1/sref48http://refhub.elsevier.com/S0142-9612(17)30773-1/sref48http://refhub.elsevier.com/S0142-9612(17)30773-1/sref49http://refhub.elsevier.com/S0142-9612(17)30773-1/sref49http://refhub.elsevier.com/S0142-9612(17)30773-1/sref49http://refhub.elsevier.com/S0142-9612(17)30773-1/sref50http://refhub.elsevier.com/S0142-9612(17)30773-1/sref50http://refhub.elsevier.com/S0142-9612(17)30773-1/sref50http://refhub.elsevier.com/S0142-9612(17)30773-1/sref50http://refhub.elsevier.com/S0142-9612(17)30773-1/sref51http://refhub.elsevier.com/S0142-9612(17)30773-1/sref51http://refhub.elsevier.com/S0142-9612(17)30773-1/sref51http://refhub.elsevier.com/S0142-9612(17)30773-1/sref51http://refhub.elsevier.com/S0142-9612(17)30773-1/sref52http://refhub.elsevier.com/S0142-9612(17)30773-1/sref52http://refhub.elsevier.com/S0142-9612(17)30773-1/sref52http://refhub.elsevier.com/S0142-9612(17)30773-1/sref52http://refhub.elsevier.com/S0142-9612(17)30773-1/sref53http://refhub.elsevier.com/S0142-9612(17)30773-1/sref53http://refhub.elsevier.com/S0142-9612(17)30773-1/sref53http://refhub.elsevier.com/S0142-9612(17)30773-1/sref53http://refhub.elsevier.com/S0142-9612(17)30773-1/sref54http://refhub.elsevier.com/S0142-9612(17)30773-1/sref54http://refhub.elsevier.com/S0142-9612(17)30773-1/sref54http://refhub.elsevier.com/S0142-9612(17)30773-1/sref54http://refhub.elsevier.com/S0142-9612(17)30773-1/sref55http://refhub.elsevier.com/S0142-9612(17)30773-1/sref55http://refhub.elsevier.com/S0142-9612(17)30773-1/sref55http://refhub.elsevier.com/S0142-9612(17)30773-1/sref55http://refhub.elsevier.com/S0142-9612(17)30773-1/sref56http://refhub.elsevier.com/S0142-9612(17)30773-1/sref56http://refhub.elsevier.com/S0142-9612(17)30773-1/sref56http://refhub.elsevier.com/S0142-9612(17)30773-1/sref56http://refhub.elsevier.com/S0142-9612(17)30773-1/sref57http://refhub.elsevier.com/S0142-9612(17)30773-1/sref57http://refhub.elsevier.com/S0142-9612(17)30773-1/sref57http://refhub.elsevier.com/S0142-9612(17)30773-1/sref58http://refhub.elsevier.com/S0142-9612(17)30773-1/sref58http://refhub.elsevier.com/S0142-9612(17)30773-1/sref58http://refhub.elsevier.com/S0142-9612(17)30773-1/sref59http://refhub.elsevier.com/S0142-9612(17)30773-1/sref59http://refhub.elsevier.com/S0142-9612(17)30773-1/sref59http://refhub.elsevier.com/S0142-9612(17)30773-1/sref60http://refhub.elsevier.com/S0142-9612(17)30773-1/sref60http://refhub.elsevier.com/S0142-9612(17)30773-1/sref60http://refhub.elsevier.com/S0142-9612(17)30773-1/sref60http://refhub.elsevier.com/S0142-9612(17)30773-1/sref61http://refhub.elsevier.com/S0142-9612(17)30773-1/sref61http://refhub.elsevier.com/S0142-9612(17)30773-1/sref61http://refhub.elsevier.com/S0142-9612(17)30773-1/sref61http://refhub.elsevier.com/S0142-9612(17)30773-1/sref62http://refhub.elsevier.com/S0142-9612(17)30773-1/sref62http://refhub.elsevier.com/S0142-9612(17)30773-1/sref62


[63] L.A. Flanagan, et al., Neurite branching on deformable substrates, Neuroreport13 (18) (2002) 2411e2415.

[64] D.K. Cullen, et al., Neural tissue engineering and biohybridized microsystemsfor neurobiological investigation in vitro (Part 1), Crit. Rev. Biomed. Eng. 39 (3)(2011) 201e240.

[65] A. Verkhratsky, Physiology of neuronaleglial networking, Neurochem. Int. 57(4) (2010) 332e343.

[66] P.M.D. Watson, et al., Bioengineered 3D glial cell culture systems and appli-cations for neurodegeneration and neuroinflammation, SLAS Discov. 22 (5)(2017) 583e601.

[67] I. Smith, et al., Neuronal-glial populations form functional networks in abiocompatible 3D scaffold, Neurosci. Lett. 609 (2015) 198e202.

[68] T.B. Puschmann, et al., Bioactive 3D cell culture system minimizes cellularstress and maintains the in vivo-like morphological complexity of astroglialcells, Glia 61 (3) (2013) 432e440.

[69] S. Balasubramanian, et al., Three-Dimensional environment sustainsmorphological heterogeneity and promotes phenotypic progression duringastrocyte development, Tissue Eng. Part A 22 (11e12) (2016) 885e898.

[70] A.L. Placone, et al., Human astrocytes develop physiological morphology andremain quiescent in a novel 3D matrix, Biomaterials 42 (2015) 134e143.

[71] V.I. Scanga, et al., Biomaterials for neural-tissue engineering d chitosansupports the survival, migration, and differentiation of adult-derived neuralstem and progenitor cells, Can. J. Chem. 88 (3) (2010) 277e287.

[72] M. Prasitsilp, et al., Cellular responses to chitosan in vitro: the importance ofdeacetylation, J. Mater Sci. Mater Med. 11 (12) (2000) 773e778.

[73] Dabiri, Seyed Mohammad Hossein, et al., New in-situ synthetized hydrogelcomposite based on alginate and brushite as a potential pH sensitive drugdelivery system, Carbohydr. Polym. 177 (2017) 324e333.

[74] Seung G. Lee, et al., Molecular dynamics simulation study of P (VP-co-HEMA)hydrogels: effect of water content on equilibrium structures and mechanicalproperties, Biomaterials 30.30 (2009) 6130e6141.

[75] G. Palazzolo, et al., Ultrasoft alginate hydrogels support long-term three-dimensional functional neuronal networks, Tissue Eng. Part A 21 (15e16)(2015) 2177e2185.

[76] M. Matyash, et al., Novel soft alginate hydrogel strongly supports neuritegrowth and protects neurons against oxidative stress, Tissue Eng. Part A 18(1e2) (2012) 55e66.

[77] C. Agulhon, et al., What is the role of astrocyte calcium in neurophysiology?Neuron 59 (6) (2008) 932e946.

[78] S.S. Lim, D.C. Foo, Simulation and scale-up study for a chitosaneTiO2 nano-tubes scaffold production, Food Bioprod. Process. 106 (2017) 108e116.

[79] I. Hamed, et al., Industrial applications of crustacean by-products (chitin,chitosan, and chitooligosaccharides): a review, Trend. Food Sci. Technol. 48(2016) 40e50.

[80] N. Garcia-Giralt, et al., Chitosan microparticles for “in vitro” 3D culture ofhuman chondrocytes 3 (2013) 6362e6368.

http://refhub.elsevier.com/S0142-9612(17)30773-1/sref63http://refhub.elsevier.com/S0142-9612(17)30773-1/sref63http://refhub.elsevier.com/S0142-9612(17)30773-1/sref63http://refhub.elsevier.com/S0142-9612(17)30773-1/sref64http://refhub.elsevier.com/S0142-9612(17)30773-1/sref64http://refhub.elsevier.com/S0142-9612(17)30773-1/sref64http://refhub.elsevier.com/S0142-9612(17)30773-1/sref64http://refhub.elsevier.com/S0142-9612(17)30773-1/sref65http://refhub.elsevier.com/S0142-9612(17)30773-1/sref65http://refhub.elsevier.com/S0142-9612(17)30773-1/sref65http://refhub.elsevier.com/S0142-9612(17)30773-1/sref65http://refhub.elsevier.com/S0142-9612(17)30773-1/sref66http://refhub.elsevier.com/S0142-9612(17)30773-1/sref66http://refhub.elsevier.com/S0142-9612(17)30773-1/sref66http://refhub.elsevier.com/S0142-9612(17)30773-1/sref66http://refhub.elsevier.com/S0142-9612(17)30773-1/sref67http://refhub.elsevier.com/S0142-9612(17)30773-1/sref67http://refhub.elsevier.com/S0142-9612(17)30773-1/sref67http://refhub.elsevier.com/S0142-9612(17)30773-1/sref68http://refhub.elsevier.com/S0142-9612(17)30773-1/sref68http://refhub.elsevier.com/S0142-9612(17)30773-1/sref68http://refhub.elsevier.com/S0142-9612(17)30773-1/sref68http://refhub.elsevier.com/S0142-9612(17)30773-1/sref69http://refhub.elsevier.com/S0142-9612(17)30773-1/sref69http://refhub.elsevier.com/S0142-9612(17)30773-1/sref69http://refhub.elsevier.com/S0142-9612(17)30773-1/sref69http://refhub.elsevier.com/S0142-9612(17)30773-1/sref69http://refhub.elsevier.com/S0142-9612(17)30773-1/sref70http://refhub.elsevier.com/S0142-9612(17)30773-1/sref70http://refhub.elsevier.com/S0142-9612(17)30773-1/sref70http://refhub.elsevier.com/S0142-9612(17)30773-1/sref71http://refhub.elsevier.com/S0142-9612(17)30773-1/sref71http://refhub.elsevier.com/S0142-9612(17)30773-1/sref71http://refhub.elsevier.com/S0142-9612(17)30773-1/sref71http://refhub.elsevier.com/S0142-9612(17)30773-1/sref71http://refhub.elsevier.com/S0142-9612(17)30773-1/sref72http://refhub.elsevier.com/S0142-9612(17)30773-1/sref72http://refhub.elsevier.com/S0142-9612(17)30773-1/sref72http://refhub.elsevier.com/S0142-9612(17)30773-1/sref73http://refhub.elsevier.com/S0142-9612(17)30773-1/sref73http://refhub.elsevier.com/S0142-9612(17)30773-1/sref73http://refhub.elsevier.com/S0142-9612(17)30773-1/sref73http://refhub.elsevier.com/S0142-9612(17)30773-1/sref74http://refhub.elsevier.com/S0142-9612(17)30773-1/sref74http://refhub.elsevier.com/S0142-9612(17)30773-1/sref74http://refhub.elsevier.com/S0142-9612(17)30773-1/sref74http://refhub.elsevier.com/S0142-9612(17)30773-1/sref75http://refhub.elsevier.com/S0142-9612(17)30773-1/sref75http://refhub.elsevier.com/S0142-9612(17)30773-1/sref75http://refhub.elsevier.com/S0142-9612(17)30773-1/sref75http://refhub.elsevier.com/S0142-9612(17)30773-1/sref75http://refhub.elsevier.com/S0142-9612(17)30773-1/sref76http://refhub.elsevier.com/S0142-9612(17)30773-1/sref76http://refhub.elsevier.com/S0142-9612(17)30773-1/sref76http://refhub.elsevier.com/S0142-9612(17)30773-1/sref76http://refhub.elsevier.com/S0142-9612(17)30773-1/sref76http://refhub.elsevier.com/S0142-9612(17)30773-1/sref77http://refhub.elsevier.com/S0142-9612(17)30773-1/sref77http://refhub.elsevier.com/S0142-9612(17)30773-1/sref77http://refhub.elsevier.com/S0142-9612(17)30773-1/sref78http://refhub.elsevier.com/S0142-9612(17)30773-1/sref78http://refhub.elsevier.com/S0142-9612(17)30773-1/sref78http://refhub.elsevier.com/S0142-9612(17)30773-1/sref78http://refhub.elsevier.com/S0142-9612(17)30773-1/sref79http://refhub.elsevier.com/S0142-9612(17)30773-1/sref79http://refhub.elsevier.com/S0142-9612(17)30773-1/sref79http://refhub.elsevier.com/S0142-9612(17)30773-1/sref79http://refhub.elsevier.com/S0142-9612(17)30773-1/sref80http://refhub.elsevier.com/S0142-9612(17)30773-1/sref80http://refhub.elsevier.com/S0142-9612(17)30773-1/sref80

Soft chitosan microbeads scaffold for 3D functional neuronal networks1. Introduction2. Materials and methods2.1. Preparation of chitosan films2.2. Preparation of chitosan microbeads2.3. Characterization of chitosan microbeads2.3.1. Water content2.3.2. Atomic force microscopy (AFM)

2.4. Cell preparation2.4.1. Preparation of 2D networks on CHI films and 3D networks on CHI microbeads

2.5. Morphological characterization of 3D neuronal networks by transmission electron microscopy2.6. Morphological characterization of neuronal networks by immunocytochemistry2.6.1. Optical microscopy and confocal imaging

2.7. MEA recording and analysis2.7.1. Data and statistical analysis

3. Results3.1. Preparation and characterization of CHI microbeads3.2. Preparation and characterization of 2D neuronal networks on CHI films3.3. Preparation and characterization of 3D neuronal networks on CHI microbeads3.4. Functional characterization of 3D networks

4. Discussion4.1. Characterization of CHI microbeads4.2. Characterization of 2D neuronal networks on CHI films4.3. Characterization of 3D neuronal networks on CHI microbeads4.4. Functional characterization of 3D neuronal networks on CHI microbeads

5. ConclusionsFundingAcknowledgementsAppendix A. Supplementary dataReferences

Date post:	29-Jan-2021
Category:	Documents
Upload:	others
View:	6 times
Download:	0 times

Soft chitosan microbeads scaffold for 3D functional ......Soft chitosan microbeads scaffold for 3D...

Documents